Additivity of masking in normally hearing and hearing-impaired subjects Andrew J. Oxenham and Brian C. J. Moore Department of Experimental Psych• logy, University of Cambridge, Downing Street, Cambridge CB2 3EB, England (Received 25 October 1994; revsed 13 April 1995; accepted 28 April 1995) The effects of combining two {:qually effective maskers werestudied in normally hearing and elderly hearing-impaired subject:.. Theadditivity of nonsimultaneous masking was investigated by measuring thresholds for a brief4-kHz signal in thepresence of a broadband-noise forward masker, a backward masker, and a combi nation of' both. Forthenormally hearing subjects, combining two equally effective nonsimultaneot s maskers resulted in up to a 15-dB greater increase in threshold than the 3 dB predicted by air energy-summation model("excess masking"). However, the hearing-impaired subjects show•d little or no excess masking. The difference between the two groups is consistent with a theory linking excess masking to the compressive transfer function measured on thebasilar membra ae(BM). In thehearing-impaired subjects thetransfer function is more linear, accounting for the la :k of excess masking. Theadditivity of simultaneous masking was investigated by measuring thresh •ldsfor a 100-ms 4-kHz signal in the presence of either a 400-ms broadband noise masker or a 400 combining two equally effectiw masking (3 to 4 dB) wassimilar based on the use of different der( peripheral compression may unc unlikely that in simultaneous ma two maskers are compressed ind that BM nonlinearity may under growthof forwardmasking, as • the masker frequencies are well mssinusoidal masker at thesame frequency asthesignal, and then maskers, a noise and a tone. The maximum amount of excess across the two groups of subjects, consistent with an explanation ction cues for thetonal andnoise maskers. It is argued that,while erlie excess masking for pairs of nonsimultaneous maskers, it is •king, where the maskers areclose in frequency to the signal, the vidually before theireffects arecombined. It is further suggested ie the effects of the upward spread of masking and the nonlinear ell asaccounting f•r the additivity of simultaneous masking when )elow that of the signal. ¸ 1995 Acoustical Society of America. PACS numbers: 43.66.Dc, 43.6{.Ba, 43.66.Sr, 43.66.Mk INTRODUCTION The question of how the effects of t• o or more maskers combine has been an area of interest for come time. In nor.- mally hearing subjects, the additivity o• nonsimultaneous maskiug (Elliott, 1969; Wilson andCarhart, 1971; Robinson and Pollack, 1973;Penner, 1980;Cokely and Humes,1993; Oxenham and Moore, 1994) and the additivity of simulta-- neous masking (Green, 1967; Zwicker and Herla, 1975:; Lutfi, 1983; Moore, 1985; Humeset al., 1992b) have been systematically investigated. The additivty of masking in hearing-impairedsubjects has received less attention. Jesteadt (1983) compared the performance: of normally hear- ingand hearing-impaired subjects using a sinusoidal forward masker, at the same frequency as tl•e signal, and a broadband-noise simultaneous masker. H• found that com-- bining these two maskers led to a smaller increase in thresh-- olds for the hearing-impaired thanfor the: normally hearing subjects. Lutfi (1987) compared the perfor mance of normally hearing and hearing-impaired subjects t sing two simulta-. neous narrow-band noise maskers centered on either side of the signal frequency. Lutfi alsofoundl•rge differences in performance between the two groups a•d proposed some possible explanations which are discusse•l in Sec. II of this paper. We know of no studies whichinve •tigate the additiw ity of purelynonsimultaneous maskers in the hearing im- paired.. Some theories of the additivityof masking have dealt with the effectsof nonsimultaneous maskers (Penner,1980; Penner and Shiffrin, 1980; Oxenham and More, 1994), an- other has treated simultaneous masking effects (Lutfi, 1983), while another attempts to provide a method of accounting for both situations (Humesand Jesteadt, 1989). All these theo- ries rely on the assumption that the individual stimuli (maskers and signal) are processed independently before their effects are combined. It is this assumption, together with the possible underlying physiological mechanisms, which we investigate in this paper. We first address issues relating to nonsimultaneous masking, andreport results from an experiment which investigates whether the additivityof nonsimultaneous masking maybe described in terms of per- ipheral auditorynonlinearities by comparing the perfor- manceof three normallyhearingsubjects with that of three subjects with cochlear hearing impairment. I. NONSIMULTANEOUS MASKING The most: common way to investigate the additivity of nonsimultaneous masking has been by studying theeffects of combininga forward and a backwardmasker.Considera forward masker and a backward masker whichindividually produce an equal amount of masking for a givensignal. An energy-summation model of masking predictsa 3-dB in- 1921 J. Acoust. Soc.Am. 98 (4), October 19½ 5 0001-4966/95/98(4)/1921/14/$6.00 ¸ 1995Acoustical Society of America 1921